论文部分内容阅读
面向CPU-GPU异构计算机体系,基于离散元法,完成了散体物料运输过程中紧急刹车的数值模拟,得到了13.01的平均加速比,结果表明,利用GPU进行DEM加速计算不仅可行,而且比CPU的串行求解的加速效果更明显。研究了刹车过程中不同区域颗粒的运动状态以及它们在不同时刻的平均运动速度,并据此来表征颗粒运动的剧烈程度。分析了不同刹车加速度下颗粒的运动状态以及对车厢前壁和车厢底板所受作用力。结果表明:同一刹车加速度下越靠近车厢尾部颗粒运动越剧烈,同一区域中的颗粒运动平均速度随着刹车加速度的增大而增大;随着刹车加速度的增大,同一区域中颗粒的平均速度也越快达到峰值并越先下降并趋于平稳,且作用于车厢壁上的动压力也在增大;加速度越大压力峰值越先到来,峰值也越大,这说明紧急刹车过程中,较大刹车加速度严重地影响了车厢压力和车内货物的分布情况。
For the CPU-GPU heterogeneous computer system, based on the discrete element method, the numerical simulation of the emergency brake in the process of bulk materials transportation is completed and the average speedup of 13.01 is obtained. The results show that it is not only feasible to use GPU to accelerate DEM calculation, The acceleration effect of CPU serial solution is more obvious. Studied the movement of particles in different regions of the brake process and their average velocity at different times, and thus to characterize the severity of particle movement. The movement of particles under different brake accelerations and the forces acting on the front wall and the floor of the carriage were analyzed. The results show that the more severe the particle movement is, the more the average particle velocity in the same region increases with the increase of brake acceleration. With the increase of brake acceleration, the average particle velocity in the same region also increases The sooner the peak value is reached and the first one is going to descend and the second one is steady. The dynamic pressure acting on the wall of the car body also increases. The larger the acceleration is, the higher the pressure peak value comes first and the larger the peak value. This shows that in the emergency braking process, Brake acceleration has a serious impact on the cabin pressure and the distribution of the goods inside the car.